Type in a word and get a bunch of youtube videos paused at a timestamp where the word appears in the captions.
The project is on free hosting plan so it can be slow on first requests.
The database on free plan can store limited number of rows so for now service is working fine on simple, common English words.
Explore :)
That’s pretty cool, how did you collect all the caption data?
With youtube API. It provide captions already split, with timestamps. I store them in db and make search with ElasticSearch.